An Original Contribution to the Modelling of Ship Collision Risk in Confined Waterways with Application to the Lagos Port Approach Channel

Ship collision risk in confined waterway environments is influenced by the complex interaction of traffic density, channel geometry, environmental conditions, and human navigational decision-making, creating a modelling challenge that existing risk quantification frameworks address only partially. This study makes an original contribution to ship collision risk modelling methodology through the development of a novel probabilistic risk model specifically designed for confined port approach channels, demonstrated through application to the Lagos Port approach channel. The model integrates three analytical components: a kinematic collision candidate detection algorithm calibrated to the traffic patterns observed in the Lagos Vessel Traffic Service data; a Bayesian network model of human navigational decision quality incorporating seafarer experience, workload, and channel complexity variables; and a consequence assessment module estimating collision damage severity using a structural crashworthiness model derived from the characteristics of vessels using the channel. The integrated model was validated against historical near-miss and collision incident data for the Lagos port approach over a ten-year period. Application of the calibrated model reveals that collision risk in the Lagos approach channel is approximately three times higher than in equivalent European port approaches of similar traffic density, primarily attributable to communication and seamanship factor contributions in the Bayesian network. The theoretical contribution lies in the explicit integration of human factor modelling within a probabilistic collision risk framework at the confined waterway level. Keywords: ship collision risk, probabilistic modelling, Bayesian network, Lagos port approach, confined waterways